1. Field of the Invention
The present invention relates to data analysis of semiconductor tests using a computer and, more particularly to a technique for a software program executed by a computer that combines (1) the measuring results obtained as a combination of arrays from a measuring apparatus and (2) the measuring results obtained as scalar representation data from the measuring apparatus and analyzes thereof.
2. Discussion of the Background Art
When an IC (integrated circuit), whether in a die on a wafer or a package form, referred to hereinafter as measured elements and DUT (device under test), are measured a variety of parameters are set and data are collected. For example, in the semiconductor measuring system described in Japanese Patent Application Publication No. H11[1999]-163,063, a variety of parameters are provided for a test spec 7d and vast amounts of measured data are obtained and evaluated. A combined Agilent 4072 Semiconductor Parameter Tester manufactured by Agilent Technologies Inc. and Semiconductor Process Evaluation Core Software (SPECS) manufactured by the same company or the Agilent 4155 Semiconductor Parameter Analyzer manufactured by Agilent Technologies Inc. and the like may be used for these measurements.
When the current/voltage (I-V) characteristics are measured relative to each of the DUTs under specific conditions using measuring apparatuses, the data are obtained as multiple combinations of arrays that indicate the voltage value relative to the current value under one or more conditions for each DUT and are stored in a file in a format desired by the user. The stored data are calculated using a dedicated data analysis program or Excel® or other all-purpose spreadsheet program. As a result, VTH and other feature values are extracted from the respective array data.
However, when processing was carried out for the obtained array data using a dedicated data analysis program, complicated operations using the array data were needed such that the program had to be revised to handle the quantity of data being processed. Also, the cost of the program increased when a dedicated data analysis program was provided that could handle the data flexibly such that the data were not controlled by the size of the array and the number of combinations of the arrays. Additionally, when the need arose to revise an algorithm used for extracting a feature value during analysis, recompiling had to be carried out and it could not be flexibly revised.
On the other hand, when arithmetic operations were carried out using an all-purpose spreadsheet program, feature value extraction from the multiple array data such as the aforementioned I-V characteristics were executed for each individual spreadsheet. Also, the extracted feature values were inserted in a two-dimensional table, that is, the main spreadsheet which had been displayed for each DUT along with the results of the other measurement items, and were evaluated. However, inserting the feature values obtained for each of the individual spreadsheets mentioned above in a cell on the main spreadsheet is a manual operation which is cumbersome and time-consuming. In addition, the process of debugging between the main spreadsheet and each of the spreadsheets used for feature value extraction is not easily done without sufficient specialized knowledge. Also, when the prior-art all-purpose spreadsheet program was used, there was no function to easily display the array or perform operations for it unless expanding the array to another spreadsheet.